What a Simulator Can and Can't Teach You

Simulator time is far cheaper than track time, available in any weather and free of risk. It is tempting to treat it as a substitute for testing. It is not — and drivers who use it that way tend to arrive at the circuit with habits to unlearn. Used for what it does well, it is one of the most efficient forms of preparation there is.

What transfers well

Some things carry over from simulator to track almost unchanged, because they depend on knowledge rather than on feel.

• Circuit knowledge — the sequence of corners, their direction, gradient and camber
• Reference points — braking markers, turn-in points and the lines that link one corner to the next
• Gears and rhythm — which gear for which corner, and the flow of a lap
• Procedures — starts, pit entry and exit, switch changes, safety-car and full-course-yellow routines
• Traffic and racecraft — where passing is possible, and where a lapped car will cost the most time

A driver who has done this work arrives at a new circuit already knowing where it goes. The first session can be spent refining rather than exploring, which on a short test day is a significant gain.

What transfers poorly

Other things do not survive the move, and it is important to know which.

• Absolute grip — the level in the simulator is a model, and the real surface, temperature and rubber will differ
• Tyre behaviour — warm-up, overheating and degradation are among the hardest things to model accurately
• Physical sensation — sustained load through the body is absent or only suggested, and it is a large part of how a driver senses the limit
• Consequence — a mistake costs nothing, and drivers take risks they would not take in a real car
• Changing conditions — track evolution, wind and temperature through a session

None of this makes the simulator wrong. It means that exact brake points, minimum speeds and lap times should be treated as starting estimates, to be corrected within the first few laps on track.

Correlation is the whole question

The value of a simulator session depends on how well the model matches the real car at the real circuit. This is correlation, and it is measured the same way as any other comparison: overlay simulator data on track data, by distance, and look at where the two agree.

Typically the match is good on the straights and in high-speed sections, where behaviour is dominated by power, drag and downforce. It is weaker in slow corners and under combined braking and turning, where tyre behaviour dominates. Knowing where a particular model is reliable tells you which of its conclusions to act on.

A simulator is most useful to the people who know exactly where it is wrong.

Trends before numbers

For setup work, a simulator is far better at direction than at magnitude. If a change improves stability in the model, it will usually improve stability on the car; how much lap time it is worth is much less certain.

That is still valuable. Arriving at a test knowing which three directions are worth trying, and which are not, saves runs. Arriving with a setup the simulator declared fastest and expecting it to be so does not.

The risk of practising the wrong thing

Repetition builds habit whether or not the habit is correct. If the model allows a brake point the real car cannot manage, or forgives an abrupt release that the real car would punish, a driver can spend hours rehearsing a technique that has to be undone at the circuit.

The safeguard is to practise what transfers. Learn the circuit, the references and the procedures to the point where they are automatic — and treat the last few tenths as something to be found on track.

Using it well

• Set an objective for each session: a circuit, a procedure, a comparison between setup directions
• Log the session and review it as you would track data, rather than judging by lap time alone
• After each event, compare track data with the simulator and note where they differed
• Use those differences to adjust the model, or at least to adjust how far you trust it

The simulator does not replace the car, and the car cannot offer what the simulator does. Preparation in one and confirmation in the other is a better use of both than asking either to do the other's job.